Litcius/Paper detail

Privacy Enhanced Edge-AI Healthcare Devices Authentication: A Federated Learning Approach

Syed Thouheed Ahmed, A. C. Kaladevi, V Kumar, Achyut Shankar, Fayez Alqahtani

2025IEEE Transactions on Consumer Electronics21 citationsDOI

Abstract

Medical devices connected via Internet are open to cyber-attacks, hence the privacy preservation of the user and the devices plays an important role. In this paper, Medical Edge-devices privacy enhanced technique based on the federated learning principles is proposed on the distributed coordination of devices and computational system. The technique generates a dedicated token via Centralized Authentication Authority (CAA) for the active users at the local Federated Learning (FL) cum training model. Further the technique introduces dual layer validation of tokens via Authentication Authority Valuation Layer (AAVL) and Centralized Server Firewall Authentication Layer (CSFAL). The purpose of these layers is to optimize the active and inactive tokens, assigning tokens and revoking the inactive tokens from the user groups via distributed federated learning model. The technique is validated on 1500 edge devices of IoT/IoMT on 5 clusters servers of global model. The technique has demonstrated an accuracy of 96% in training and optimizing the IoMT device operations via privacy policies proposed in this paper.

Topics & Concepts

Computer scienceAuthentication (law)Health careInformation privacyComputer securityEnhanced Data Rates for GSM EvolutionComputer networkInternet privacyArtificial intelligenceEconomicsEconomic growthPrivacy-Preserving Technologies in Data
Privacy Enhanced Edge-AI Healthcare Devices Authentication: A Federated Learning Approach | Litcius